0:00

I was in Ghent at configuration

0:02

management camp earlier this week, Monday,

0:04

Tuesday. I saw most of the

0:06

usual people that I would see

0:08

there. Always nice to hang out,

0:11

see people. They had a trailer

0:13

with Belgian waffles, very good. I

0:15

only ate in that cafeteria once.

0:17

I got a little tiny lasagna

0:19

in a metal thing. It wasn't like

0:22

the best choice. No, like cafeteria

0:24

lasagna, come on, that's terrible. I

0:26

know, but you know, but you know,

0:28

it was okay. It's one of those

0:31

things. There's a category of food that

0:33

I don't really eat very much at

0:35

all. And so when I do eat it,

0:37

it's kind of novel and okay. And I,

0:39

you know, I'm not disagreeing with

0:41

the sentiment. The lasagna is

0:44

difficult. But, like, I finally got

0:46

the chance to give this talk

0:48

that I've been working on for

0:50

a little while about how much

0:52

public versus private cloud there is.

0:54

Now, people who subscribe to my

0:57

newsletter, if you go to newsletter. which

0:59

I need to get some subscribers

1:01

because I've lost a few. Maybe that's

1:03

because I've been talking about this too

1:06

much or other things, but right now

1:08

it's at 889 people. We just

1:10

need to have like maybe 1112 or

1:12

like 300 more. That'll be a fun

1:14

number to have. So you got to subscribe

1:16

to that. Now, this whole talk I

1:18

gathered together, I didn't even think

1:21

they were going to accept this talk,

1:23

but I gathered together a bunch of

1:25

charts. Just going over like where the

1:27

workloads are or or our friend Donnie

1:29

was like workloads come on old man

1:31

And I think he was right. You

1:33

can't say workloads, but I say it

1:36

anyways the applications And I think you know,

1:38

I'm not going to go over those results

1:40

They're going to post a video at some

1:42

point you can go see the stream and

1:44

scrub to find it. It's basically I don't

1:46

know 40-60 or 50-50 depending on the

1:48

charge you have but That's not

1:50

the important thing I want to

1:53

get some consulting from the two sports

1:55

people I know baseball is not y'all's

1:57

game. You're more into like, I don't

1:59

know... medicine water polo or whatever. But

2:01

I tried to take seriously, to make

2:04

a joke, sometimes you've got to take

2:06

things seriously, I tried to take seriously

2:08

this early inning situation, the first inning.

2:10

So I charted out going all the

2:12

way back to 2010, you know, for

2:14

the time I allocated to this. I

2:17

didn't have access to either Google or

2:19

Open AI deep research or whatever. So

2:21

maybe I'll use that next time to

2:23

look this up. But I found our

2:25

old friend Peter Ulander. He had a

2:27

quote when he was at cloud.com that

2:30

were in the first ending of cloud

2:32

2010. And then you can chart this

2:34

all the way through. I think they

2:36

stopped saying it. You'd think having made

2:38

these slides that would have them in

2:40

front of me to see when anything

2:43

stopped. But at some point, it did

2:45

switch over to a 10K. You know.

2:47

Well, that was that was in 2023.

2:49

You know, how do you say it?

2:51

Adam Selepski. He was saying we're in

2:53

the first few steps of a 10K.

2:55

However, in that same year, the IBM

2:58

said the cloud is probably in its

3:00

fifth inning. So IBM's like living in

3:02

the future. Yeah. I mean, so IBM's

3:04

calling for like total cloud consolidation in

3:06

20 years, whereas Selepski, like, you know,

3:08

we are. 2,000 years away. Something like

3:11

that. And then, but then, but then

3:13

this, this past month since it's February

3:15

5th, you know, Matt Garmin, AWS, CEO,

3:17

he said that we're still at the

3:19

early stages of what the business can

3:21

be. So, I think, no sports metaphor.

3:24

No, no sports metaphor. I think maybe

3:26

they have some AWS comms people and

3:28

they must listen to this podcast. They

3:30

were like, whoa, these guys, for once

3:32

they've got kind of a good point

3:34

about this ending thing. But for once

3:36

I thought I thought what I would

3:39

do. I like that they have have

3:41

listened to you know, what were you

3:43

on 500 and to know. Like five

3:45

weeks. They made a point. You know,

3:47

they just they just need recommendations. They

3:49

just wait the whole time. So I

3:52

thought I thought what I would do

3:54

is I would chart out. Let's say,

3:56

you know, let's go with, if we're

3:58

still early and we're going to stick

4:00

to a baseball game, I would take

4:02

early, I mean, I'm going to get

4:05

to the question. I would take early

4:07

to mean we're in inning two or

4:09

three. Now this is because. And I

4:11

was in, you know, Belgium, so I

4:13

felt the need to go over American

4:15

baseball stuff. It's like you got nine

4:18

innings, right? Now you can, of course,

4:20

go overtime, and the longest game ever,

4:22

I think, went 27 innings, or 26

4:24

innings, I'm sorry. And that was like

4:26

a 19-20. But that also assumes it's

4:28

a tie after nine innings. Right, exactly.

4:30

Yeah, we're not tracking for that. But

4:33

you know, all these quotes, they haven't

4:35

said anything about a tie or not.

4:37

They just strictly on the innings. What

4:39

are we tying up against? Oh, and

4:41

I haven't thought of this. Here we

4:43

go Matt, right? Workloads. Because possibly we're

4:46

at a 50-50 split between where the

4:48

workloads are, we might be tied so

4:50

far. How long is this a race

4:52

against on- How long is this race

4:54

against on- there could always be extra

4:56

innings and who knows what that means

4:59

right that's right exactly that's the variable

5:01

I think I think maybe if someone

5:03

nails quantum computing we're going to be

5:05

in extra innings I say you would

5:07

call that you would call that a

5:09

walk-off home run oh a walk-off home

5:12

run okay okay okay okay so I

5:14

thought let's just assume a nine-in-in-ing game

5:16

we're not going to go crazy right

5:18

and I wanted to chart out so

5:20

depending on which inning we're in. And

5:22

again, I take early to mean the

5:24

third inning, second or third. Because if

5:27

you're in the first inning, you're like,

5:29

I'm in the first inning, right? You're

5:31

never in the first something, you're like,

5:33

we're early. So we got to be

5:35

past the first inning and then if

5:37

you were in the fourth or fifth

5:40

inning, you would be like out of

5:42

nine I don't think in base nine,

5:44

but you'd be like we're halfway Right,

5:46

and so like it's got to be

5:48

the second or the third so I

5:50

thought what I would do is I

5:53

would go with like the third we're

5:55

in the third inning and then I

5:57

somehow have to count and so I

5:59

want as you're getting to Matt Ray

6:01

I need to figure out when this

6:03

this this this this god damn game

6:05

is over because it's been going on

6:08

for a long time. And so I

6:10

thought I would do some, you know,

6:12

liberal artsy charts where I couldn't figure

6:14

out what the kind of rate, how

6:16

long innings you usually are and what

6:18

the rate of changes as the game

6:21

goes on. Like, do the endings get

6:23

longer? Do they get faster? Right? Like

6:25

my observation with other sports is the

6:27

longer the game goes, the slower it

6:29

takes to finish because you keep taking

6:31

timeouts and you're just fucking about. So,

6:34

like, the last part of the game

6:36

takes the longest to do, kind of

6:38

like cloud migrations, where you can move

6:40

90% of it, and then it takes

6:42

you 10 years to move the other

6:44

10% of what you're doing. So no

6:47

one would really tell me, and by

6:49

that I mean the two people who

6:51

respond to this. So, what is the

6:53

rate that I should figure for innings

6:55

four? to nine when I'm doing this

6:57

because then I feel like I can

6:59

make a pretty good chart if we

7:02

if we had some I don't know

7:04

how to do a Monte Carlo simulation

7:06

but I think if we could figure

7:08

out like well you're looking at like

7:10

a 2% lengthening on average or no

7:12

actually the the innings they double or

7:15

no they they they they go faster

7:17

like what do we do in here

7:19

as far as the speed of innings?

7:21

Yeah I'm gonna go my my non-scientific

7:23

take is it's probably a wash and

7:25

they're probably all about the same because

7:28

sometimes sometimes like you know first second

7:30

inning, one team just unloads on them,

7:32

and the rest of the game is

7:34

like, let's just get this thing over

7:36

with, right? You know, it's like, oh,

7:38

they're up 10 runs. We're not going

7:41

to score 10. So you don't really

7:43

put in the effort. You're just kind

7:45

of like, yeah, you're just building up

7:47

the numbers so that you can get

7:49

more runs? Is that right? You want

7:51

to get that on your scorecard? Yeah.

7:53

Well, the beauty of the baseball metaphor

7:56

metaphor here. infinite time like you know

7:58

if you don't get three outs you

8:00

could in theory play the game forever

8:02

much I think you know cricket has

8:04

something so much so so that is

8:06

that does draw quite a a difficulty

8:09

in creating some type of like interval

8:11

around it because an interval there you

8:13

know so it's like well you probably

8:15

just like Matt was saying you have

8:17

to allude to some type of average

8:19

well you know an average inning takes

8:22

two years again cloud world right like

8:24

in if you started in 20 If

8:26

we go back to AWS, right, so

8:28

that's 15 years ago, something like that,

8:30

something on the order of like, you

8:32

know, 2010, 2015, something like that. So

8:34

if you said that started it 10

8:37

years into it, if we're still in

8:39

the third ending, so every three endings

8:41

is 10 years, so every three endings

8:43

is 10 years, you know, maybe that

8:45

could, you could say that, maybe every

8:47

three endings five years? Yeah, right, right,

8:50

right, so then, so then, what, okay,

8:52

so, I did several, I did several,

8:54

I did a chart, on various different

8:56

things. Now, I think the point that

8:58

you two are making, well, even more

9:00

or less, Matt Ray mostly said it,

9:03

if the innings are basically, I don't

9:05

want to say constant, but if it's

9:07

a wash, so you kind of count

9:09

them as the same length. What we're

9:11

looking at is based on, I think

9:13

I base cloud starting in 2006 when

9:16

Amazon S3 launched, you know, whatever, you

9:18

got to pick something. right? And I

9:20

always forget if it was S3 or

9:22

EC2 that was first or if it

9:24

was some some networking thing or what

9:26

it doesn't matter, right? So like, you

9:28

know, if we are in inning three

9:31

and we're still an inning three and

9:33

it is 2025 then you know you

9:35

can start to calculate how long these

9:37

things take and I think if you

9:39

if you use that if you use

9:41

the Matt Ray method it's pretty much

9:44

the same we're looking at 2068 to

9:46

end this game. But what is what

9:48

is the end of the game? Now

9:50

hold on, I'll get to that because

9:52

there is a consideration there. Now, the

9:54

fastest rate is what I'm going to

9:57

call the IBM 1 to 1 rate.

9:59

And the 1 to 1 model assumes

10:01

that after this initial stuff, you're just

10:03

doing one inning a year, right? Like

10:05

you just speed up, right? Now there's

10:07

another model where I call it the,

10:10

I think it's the doubling rate, and

10:12

that is like... Every year it gets,

10:14

it gets, you know, until it gets

10:16

down to one, you reduce it. So

10:18

normally it would take 10 years and

10:20

now it's five and then it's two

10:22

and a half, you know, or whatever.

10:25

So something like that. And so, you

10:27

know, I just had to come up

10:29

with something because I don't understand anything.

10:31

Maybe IBM's just counting down when it's

10:33

game over for them. Right. Well, yes.

10:35

And so now if we go with

10:38

the IBM Matt Ray method, right, right?

10:40

It becomes constant from this point on

10:42

to the future. That's what we have.

10:44

So we've got a big range. It

10:46

either comes out to 2027, we're going

10:48

to be done, or 2068. Somewhere in

10:51

that range. Now, to your point, Matt

10:53

Ray, when I was going over this,

10:55

I was looking at the Matt Garmin

10:57

quote, and I think you've identified something

10:59

that popped in my head. He says,

11:01

we're just at the early stages of

11:03

what that business can be. Now he

11:06

doesn't mean, he doesn't mean workloads, he

11:08

means revenue. He means revenue. Right so

11:10

like maybe in and I'm sure they've

11:12

done a lot of research They've got

11:14

they had they brought in the dark

11:16

the barking dogs. They did some PR

11:19

facts. They thought about customers a lot

11:21

customer first and then they had a

11:23

chart and they were like, okay, we've

11:25

done the research, the business, like we

11:27

know more ending stuff, but the total

11:29

amount of revenue we're gonna get, that's

11:32

what we're targeting. We're early in that

11:34

inning, but then the question becomes what

11:36

is the maximum revenue, right? Or, you

11:38

know, is it, what do you think,

11:40

what would, what would our friend, Jamman

11:42

Bell, say, are we looking at trillion?

11:45

you know total cloud revenues across everything

11:47

so let's say it's something around that

11:49

that okay no all together so you

11:51

basically you could say you're around 30%

11:53

on your I like I like this

11:55

because I maybe before I present this

11:57

again I need to come up with

12:00

a model where we first start with

12:02

the total revenue and I can just

12:04

pull some IDC stuff or or maybe

12:06

I can just and then we can

12:08

figure out the the the the cagger

12:10

the rates of changing the rates of

12:13

change that happened and then we can

12:15

map that into the baseball Right? And

12:17

then we can figure out, okay, so

12:19

based on the innings, we know the

12:21

end to your point, Matt, Ray, the

12:23

revenue, and we know the rate of

12:26

change, the rate of growth for that

12:28

revenue, and then we can figure out

12:30

how many more innings that we have,

12:32

and we'll calculate that pretty easily, I

12:34

think. Well, listen, I could, yeah, I

12:36

want to maybe quickly propose an alternative

12:38

way to map this for you, for

12:41

your consideration, was like, what have we

12:43

just took growth rate rate, Triple digits

12:45

are higher growth rate. You're in the

12:47

first inning, right? That's definitely the first

12:49

inning. If you're in what I would

12:51

call High double digits growth rate, maybe

12:54

80% right? Maybe that's second and third

12:56

inning. Yeah, then fourth, fifth and sixth

12:58

is sort of like starting to be

13:00

like, you know, lower double digits, right?

13:02

30% 20% 18 something along this line,

13:04

which just like kind of where I

13:07

think we are today. And so then

13:09

again back to being an untime untime

13:11

on time game. You can stay in

13:13

these innings for as long as you

13:15

want as long as... the growth rate,

13:17

you know, sort of like hovers there,

13:20

you just kind of keep staying there.

13:22

And then as you kind of get

13:24

towards the end, when you start to

13:26

drop down into the single digits, a

13:28

growth rate, maybe that's the seventh or

13:30

eighth, and maybe when you start to

13:32

flat line or go down, maybe that's

13:35

the final ninth. So that would be

13:37

a way to like to do it.

13:39

And that's when you're bringing your closer

13:41

to to shut out the competition and

13:43

lock it down. Right, that's private equity

13:45

being the closer. You bring in private

13:48

equity to close it out. It is

13:50

closing out, right? You know, private equity

13:52

is a fast ball to the head.

13:54

Yeah, it's true. Maybe that's what it

13:56

is. So, so I based on that,

13:58

maybe we'll talk about clown earnings here

14:01

in a second, but like we would

14:03

say, I think we're in the middle,

14:05

we're in like the third, the third

14:07

or third or third or fourth or

14:09

fourth. I think we've got a lot

14:11

of models that we could do here

14:14

and I think maybe before the next

14:16

time I give this talk, maybe I'll

14:18

make a transcript of what we just

14:20

talked about and I'll give it to

14:22

ChatGPT and I'll be like, maybe some

14:24

charts. Oh, let me say ChatGPT. Please

14:26

explain baseball too. Well, hey, give us

14:29

a little meta commentary behind scenes when

14:31

you're presenting this. Were people paying attention

14:33

or were they just like, I have

14:35

no idea what he's talking about. I

14:37

think you know I think people were

14:39

paying attention. I think there weren't a

14:42

lot of people in the room which

14:44

is good that there was a recording.

14:46

Because there's always a lot of people

14:48

on the internet. But it's like you're

14:50

just it's a new presentation you're taking

14:52

it on the road. You're working and

14:55

not working out the kinks. You know

14:57

you can't you don't want to take

14:59

your best material. Now now now we

15:01

did we did have we did have

15:03

our friend John Willis in the front

15:05

row. He knows but he was wearing

15:07

a New York Yankees's hat. He was

15:10

wearing a New York Yankees's hat. He's

15:12

hat. He's hat. Doesn't necessarily say a

15:14

lot because everyone wears that, but like

15:16

I know I know he knows baseball

15:18

So he had a few notes and

15:20

he was smiling a little bit. So

15:23

I think I think he was he

15:25

was trying to pay attention and calculate

15:27

it So you know that was good.

15:29

But yeah, I think you know it

15:31

like you're saying Brandon is a good

15:33

first run. I think there's some improvements.

15:36

This is like the after game interview.

15:38

I think I think it was up

15:40

there. I was given the presentation moving

15:42

through the slides and I think you

15:44

know you're always looking to improve I

15:46

think next time I'm just gonna hit

15:49

each slide and just do a little

15:51

better and you know you just get

15:53

through them and and then it ends.

15:55

I like this just quit mine and

15:57

correction that would be your post game

15:59

interview. That's a post game. What did

16:01

I call it? After game design. That

16:04

doesn't, that's not going to, it doesn't

16:06

work with your metaphor. Say that, yeah,

16:08

in the post game interview, it's like,

16:10

we gave it all had, Coach, the

16:12

slides really thought went well, you know,

16:14

we left it on the field, people

16:17

really like the baseball metaphor, but we

16:19

can improve. We'll do better next night.

16:21

Yeah, the, you know, the HDMI hookup,

16:23

flawless, and you know, sometimes you walk

16:25

a little too far and the clicker

16:27

doesn't work, but then you just come

16:30

back in, you just come back in,

16:32

it, it, it, it, it's just come

16:34

back in, it, it, it, it, it's

16:36

okay, it, it's okay, it, it, it's

16:38

okay, it, it, it, it, it, it,

16:40

it's, it, it's, it, it's, it, it's,

16:43

it's, it, it, it's, it, it's, it's,

16:45

it, it, it's, it, it I will

16:47

actually, we will conduct a proper post

16:49

game interview and we'll, we'll post it,

16:51

maybe that will be our first YouTube

16:53

short or, or, or, or, you know,

16:55

Instagram Real, that we actually, or Tiktok,

16:58

that we actually post, that would be

17:00

good. The Cotay Post game, post-resitation interview,

17:02

would be great. That would be good.

17:04

Well, you know, it's, it's quite fortuitous,

17:06

if that's the right word that we're

17:08

discussing this because there were cloud earnings,

17:11

cloud earnings that came out this week.

17:13

I wish, you know, maybe I should

17:15

have paid attention to work that into

17:17

the model. I mean, I think what

17:19

I need to do is make a,

17:21

I'm going to call it a living

17:24

model, where every time, you know, the

17:26

big threes, the growth numbers come out,

17:28

I can just plug it into the

17:30

model and we can readjust. to maybe

17:32

I have a whole newsletter based around

17:34

this, you know, which inning are we

17:36

in? Or when is this god damn

17:39

game going to end? We can readjust.

17:41

And it looks like, it looks like

17:43

the game's going to go on a

17:45

little bit longer unless AWS picks up

17:47

the slack. Is that the right baseball

17:49

terminology? Are you always picking up slack?

17:52

Okay. Because it looks like the expected

17:54

growth rates of Microsoft and Google a

17:56

little down. They're going spanning time. They're

17:58

moving the ball around showing some. hustle.

18:00

I guess you don't move the ball

18:02

around in baseball do you? You're getting

18:05

some hits. They're putting some runs on

18:07

the board. Yeah their box score is

18:09

still going to be good but not

18:11

as good as we hope for. And

18:13

I think you know I listen to

18:15

I listen to the the Ben Thompson

18:18

overview of the Google stuff. I think

18:20

he's always finding little nuggets. I think

18:22

the one nugget he found is that

18:24

their CFO said that they to summarize

18:26

it. They didn't have enough hardware. to

18:28

meet demand for AI stuff. And I

18:30

think this is maybe a broader thing,

18:33

which I don't know if all the

18:35

AI people got together after next week

18:37

and they're like, never mind a fastball

18:39

to the head, we gotta lock this

18:41

down. Like what are we gonna say

18:43

that shows that everything's okay? And maybe

18:46

what they're getting, and you know, probably

18:48

true as well, they're just like, well,

18:50

we have so much demand, we can't

18:52

even satisfy it. So Jeven's paradox. Let's

18:54

see how that works. And I don't

18:56

know, it looks like, I don't know

18:59

if, if Matt Ray DiALio has looked

19:01

at his, how do we like to

19:03

say it? Invidia? Invidia? However, yeah. And,

19:05

and like, how, how's it looking? Has

19:07

it, has it, has it gone back

19:09

up? Let's give, give us an update

19:12

from the post, so. Oh. I'm not,

19:14

I don't, I don't have any regrets.

19:16

If it drops 10% I have a

19:18

buy in, but that doesn't seem to

19:20

be happening. So I, I, so far,

19:22

so good, I'm not looking stupid. Well,

19:24

quick follow up from last week, we

19:27

did get, you know, I think we

19:29

talked about last week that hey, it

19:31

would take a little time for people

19:33

to like try to replicate it. So,

19:35

yeah, semi analysis, which is one of

19:37

the, before I mentioned Ben Thompson, one

19:40

of his favorite sites. They have come

19:42

out with a pretty lengthy write-up. And

19:44

so to just summarize it, it's something

19:46

to the, basically says that like, hey,

19:48

it. probably took them billions of dollars

19:50

too. They just don't want to release

19:53

the fact that they've got all these

19:55

chips that they're not supposed to have.

19:57

So that's, they did not, if you

19:59

will, release that part. So the part

20:01

they released was only the kind of

20:03

like, oh, we did it all for

20:05

five million. It's like, we did it

20:08

for five million after spending, maybe not

20:10

billions, that's a little bit, hundreds of

20:12

million, something of the order of magnitude

20:14

that other cloud vendors had, or like,

20:16

again just just one data point to

20:18

it so it's like again I think

20:21

we have to let's I think as

20:23

we talked about last week like no

20:25

one knows anything so let's just slowly

20:27

watch like you know because it will

20:29

either be replicated or it won't I

20:31

think some of the other technical analysis

20:34

has said like they have done a

20:36

lot of efficiency improvements and the stuff

20:38

they've open-source you know people are like

20:40

digging through that and they're you know

20:42

postulating like this is going to allow

20:44

us to go even do even better

20:47

on latest gen hardware on latest gen

20:49

hardware And, you know, which, you know,

20:51

you're like, oh, so that's, who's that

20:53

good for? It's like, that's still kind

20:55

of tops off invidia. It blunts the

20:57

chart, the need to charge ahead with

20:59

always the latest. It does mean if

21:02

you can afford the latest, you can

21:04

get the most performance out of it.

21:06

I mean, definitely that's there. consumer base.

21:08

So, so, you know, Jeven's paradox is

21:10

going to be thrown around a lot.

21:12

And, you know, there's something to that.

21:15

It's something that's a price. And something

21:17

becomes more ubiquitous. Yeah, making the case

21:19

for Jeven's paradox. I was like, hey,

21:21

now that everybody can do it. Everybody

21:23

needs a little bit more. But I,

21:25

my, my, my, my investment theory to

21:28

getting out of it. It was like,

21:30

yeah. That's true. But it also means

21:32

like all the competitors, like all the

21:34

competitors can catch up, can catch up

21:36

and, can catch up and enjoy that

21:38

bigger pie. I think I think you

21:41

know you you went over your spreadsheet

21:43

last week which is basically just like

21:45

I've made so much off of this

21:47

that like whatever whatever the risk is

21:49

is not worth it right so like

21:51

now yeah I was like you know

21:53

it's it's time to take your way

21:56

again we should revisit rule one no

21:58

one does anything including Matt right do

22:00

not do not do not trade against

22:02

do not trade on that raise advice

22:04

he probably does he probably has no

22:06

idea what he's talking about well let

22:09

let me let me try we don't

22:11

know I want to you know in

22:13

my head this is the way I've

22:15

I filed away and summarize this whole

22:17

big incident and let me tell you

22:19

People at config management came lots and

22:22

were just like that was the crazy

22:24

shit I've ever seen they're just like

22:26

all over this is a big deal

22:28

and I think I think maybe this

22:30

is one of the few times when

22:32

the almighty shareholder Just freaked the fuck

22:34

out and just should have like gone

22:37

back to bed and and like things

22:39

just went wacky it didn't make sense

22:41

and as a technologist, right like you're

22:43

like yeah, that's great. That's how technology

22:45

works and now we're going to use

22:47

a lot more of it and you

22:50

still need all the stuff Right? And

22:52

so it seems like maybe, you know,

22:54

because I think the key explanation, I

22:56

don't know if the semi-people covered this

22:58

or if it was just Ben Thompson's

23:00

recollections, is like, well, they published this

23:03

back in December in some PDF, and

23:05

it's only since they actually released an

23:07

app, and then the hype cycle, the

23:09

conversation cycle began that this was affected.

23:11

So, you know, what that means is

23:13

that it's just because... Normies if you

23:16

will we're paying attention and freak out

23:18

and sued. So I think I think

23:20

it's fine it's just sort of like

23:22

maybe there is not maybe this is

23:24

more of a calms issue then it's

23:26

really like any sort of like actual

23:28

product issue that you have and you

23:31

know it just means you can do

23:33

more which maybe does this bring us

23:35

back to a little bit around the

23:37

Google stuff that was kind of so

23:39

well documented by Ben Thompson and at

23:41

least I guess the CFO said something

23:44

like. They were, you know, if you

23:46

will, demand, I guess demand constraint or

23:48

fulfillment constraint, right? They. plenty of demand

23:50

they just couldn't fulfill it and it's

23:52

like and so I guess the numbers

23:54

were they they came in at this

23:57

is GCP came in 11.96 billion against

23:59

what they were projected to be 12.19

24:01

billion so again we should all have

24:03

these problems but like it's I don't

24:05

know it comes back to me I'm

24:07

just sort of like like I don't

24:10

know there's a there's a note here

24:12

on another subject later it talks about

24:14

like toxic positivity and I just I

24:16

don't know if there's like a corollary

24:18

to like earnings cost like earnings costs

24:20

it's like Really you didn't make this

24:22

money because you didn't have enough AI

24:25

hardware and that's where all this money

24:27

was left It's like it feels a

24:29

little bit like that just feels very

24:31

convenient You know I don't I don't

24:33

know what I hear that I just

24:35

I don't know what to say I

24:38

don't want to say like that's a

24:40

lie, but I just I just feel

24:42

like well I mean more truth there's

24:44

there's more to be said on that

24:46

subject that they're not when I was

24:48

listening to it I was listening to

24:51

have like Well, you think you would

24:53

have heard people saying that, like, I

24:55

can't, like, I try to run some

24:57

query against Gemini and it says no,

24:59

right? Like, I, like, because that's what

25:01

it would imply, right? Or, or I

25:03

guess, I guess what it could imply,

25:06

it could be more complicated, it could

25:08

say, like, well, when we put together

25:10

the sales plans for this quarter, we

25:12

told them the plan could have been,

25:14

you know, five percent higher per rep,

25:16

but we know we didn't, we have

25:19

capacity. So we have capacity. So we're

25:21

like, so we're like, we're like, don't

25:23

sell that. Like, you go out there,

25:25

you just make your number and don't

25:27

sell above it because we can't fulfill

25:29

capacity. And so, you know, the assumption

25:32

being that, like, you can just tell

25:34

your reps a number and they'll do

25:36

it. Right. There's not some, you know,

25:38

a market demand that drives that. But

25:40

it does, there's not some, you know,

25:42

a market demand that drives that drives

25:45

that. But it's not some, you know,

25:47

a market demand that drives. But then

25:49

they're like, well. There just wasn't enough

25:51

flour to make the tortillas people wanted

25:53

to buy that's why people would go

25:55

in there and the shells would be

25:57

empty. Sorry. Like what are you going

26:00

to do? Terror. Right. But I think

26:02

what you hit on there would be

26:04

like, we would like, like, maybe tortillas

26:06

are a better example. Like, eggs, like

26:08

you go to Costco, United States, and

26:10

now there, there are no eggs. So

26:13

it's very clear that there's not enough

26:15

eggs to fill demand, at least for

26:17

Costco, right? So, whereas I don't, I

26:19

don't, I haven't, I haven't done this.

26:21

So this is always a little dangerous

26:23

to say this. We haven't a little

26:26

dangerous to say this. We haven't, we

26:28

haven't, we haven't, we haven't, we haven't,

26:30

we haven't, we haven't, we haven't, we

26:32

haven't, we haven't, we haven't, we haven't,

26:34

we haven't, we haven't, we haven't, we

26:36

haven't, we haven't, we haven't, we haven't,

26:39

we haven't, we haven't, we haven't, we

26:41

haven't, we haven't, And I was like,

26:43

I need to spin up some AI

26:45

stuff. I feel like it would work.

26:47

Now maybe you would say that like,

26:49

I feel like maybe maybe you'd say

26:51

like, well, I'm not like trying to

26:54

buy enough, which is true. Like I

26:56

can't like, you know, try to like

26:58

spin up a $30,000 cluster or something.

27:00

But but it just feels like it's

27:02

there. That's why I'm always feel like

27:04

there's this weird, like the CFO just

27:07

makes it sound like just what you

27:09

said. it would have worked right I

27:11

would have I have I have one

27:13

more scenario it could be that there

27:15

were several key deals where someone wanted

27:17

to buy a big block of a

27:20

lot of AI yeah and they were

27:22

like can't do it and so therefore

27:24

they missed out on like huge deals

27:26

because they couldn't guarantee like ELA if

27:28

you well the huge okay that's right

27:30

right right and so like some giant

27:32

bank was like hey we need 200

27:35

million of Jim and I and and

27:37

they're like oh Sorry, even Costco doesn't

27:39

have that many eggs. Okay, that would

27:41

be a good explanation. I could see

27:43

that, but which means like all of,

27:45

you know, all of us small fries

27:48

with their 30K deals can fit in,

27:50

no problem. But if you're hunting whales,

27:52

you know, you can't land those. But

27:54

it's a classic bin packing problem and,

27:56

you know, your eggs just so they're

27:58

packing their bins, but they're saying we

28:01

need bigger bins. Well quick aside. I

28:03

just wonder you know Kote you you

28:05

mentioned somewhere on here that You're using

28:07

open AI canvas and you like it,

28:09

which I do too I was wondering

28:11

on the Gemini side. I recently because

28:14

I like open AI stuff so much

28:16

I went ahead and did a month's

28:18

subscription of Gemini thinking like it would

28:20

be like canvas but actually in Google

28:22

docs which is great then I don't

28:24

have to like cut and paste right

28:26

between the two like get it the

28:29

way I want and but my experience

28:31

with Gemini very underwhelming like it just

28:33

did not it just does not do

28:35

like I don't know I just feel

28:37

like the open AI canvas like I'm

28:39

just like working on the dock and

28:42

chatting and it's like It's figuring it

28:44

kind of knows what I want to

28:46

be done whereas like the Google things

28:48

like oh we can't sorry we can't

28:50

update your document like they'll tell you

28:52

on the side like here's the rewrote

28:55

version and then they'll paste it there's

28:57

a little like little like little clip

28:59

like they'll just paste it there's a

29:01

little like little clip like but they'll

29:03

just paste it right in there's a

29:05

little like little clip like but they'll

29:08

just paste it right to get your

29:10

Well I haven't used Jim and I

29:12

for two months I think. Maybe even

29:14

even longer because I mean I gave

29:16

it I think I started using it

29:18

maybe in the spring of past year

29:20

so I gave it lots of opportunities

29:23

and it just like like everyone has

29:25

the same reaction I asked people this

29:27

all the time is it's just I

29:29

think I think the word you might

29:31

want to use is milk toast. It's

29:33

just very boring. and you know kind

29:36

of weak sauce I it's it seems

29:38

it's weird right like it's it's you

29:40

know the only bigger disappointment in the

29:42

AI space is Apple intelligence where you

29:44

just like you know I guess they

29:46

were busy replicating Evite instead of like

29:49

working on how they're gonna do this

29:51

but like it it just like yeah

29:53

yeah I expect I expected so much

29:55

and like everything you said and it

29:57

just doesn't really satisfy and man Gemini

29:59

is the worst at playing D&D like

30:01

and and I I think I think

30:04

it's also like my theory and maybe

30:06

just like what if I can find

30:08

you to just document editing because I

30:10

thought for sure it would be the

30:12

best because of the docs where not

30:14

I feel like it's not even well

30:17

I mean you've got you've got two

30:19

things there right I don't I don't

30:21

think anyone has figured out the right

30:23

user experience for document editing right so

30:25

so this is a little bit of

30:27

a preview you know because I and

30:30

I don't want to come up with

30:32

a new one it might be in

30:34

recommendations but I haven't using canvas a

30:36

lot recently and I don't like their

30:38

UI thing, but it's okay. One thing

30:40

you notice is it's very important that

30:43

they have versioning in it. So you

30:45

can go back and look at the

30:47

history, because if you don't realize that,

30:49

the way the way the way canvas

30:51

and open AI works is you write

30:53

a document, or you write something, and

30:55

it opens up into an editable window,

30:58

and then you can highlight things and

31:00

tell it what you want, you can

31:02

do the whole thing, and then it

31:04

has, this part's pretty cool. Good job,

31:06

HDML nerds, but like you go in

31:08

there and it highlights the thing that

31:11

it's going to fix and you can

31:13

watch it kind of like spelling out,

31:15

putting the new characters for the thing.

31:17

And as a writer, at first that's

31:19

alarming because you're like, whoa, what if

31:21

I don't like it, right? Like where's

31:24

my text going? But you know, you

31:26

can go back and do things. But

31:28

it's still, there's still something that doesn't

31:30

quite feel right. But I think yes,

31:32

what you want is like I'm in

31:34

because eventually. They've got it right there.

31:37

They've got a copy paste button, and

31:39

you're just going to put it in

31:41

Google Docs, right? That's where it's going

31:43

to go. And so like, I don't

31:45

even know what the experience would be

31:47

like in Google Docs. And then to

31:49

be comprehensive, you know, you use the

31:52

Apple intelligence writing tools, and they're just

31:54

obviously some sort of. H.P. Lovecraft jibbering

31:56

beast out in space. And like, you

31:58

highlight something and whatever that, you know,

32:00

one of the five types of windows

32:02

that you have on any OS, some

32:05

strange window pops up. And it tells

32:07

you the text that it's going to

32:09

replace it. And you can. and then

32:11

it'll just dump there and replace it.

32:13

You're kind of getting to, I guess,

32:15

what I was going on is like,

32:18

it really, you know, you, I really

32:20

matter. Because I agree, I think Apple

32:22

Intelligence is kind of the worst because

32:24

it's just like, you kind of cut,

32:26

you're kind of highlighting things and it's

32:28

just doing things randomly and it's like,

32:30

there isn't any way to like go

32:33

back and version. There's no history, like,

32:35

yes, that one's sort of like the

32:37

worst. Canvas to me is probably the

32:39

best because it will, it'll let you

32:41

go back and forth and it will

32:43

kind of like edit the document for

32:46

you and give you the versions. So

32:48

I think that's really good. And then

32:50

of course the gold standard which is,

32:52

this is not, this is cheating, is

32:54

cursor. Cursor for coding is like this

32:56

works exactly the way you want. It's

32:59

like, oh, do you think I should

33:01

do all that? Makes all the changes,

33:03

highlights all the diffs, let you go

33:05

through. the coding. That's what I was

33:07

going to get to is I guess

33:09

what what I don't know if I

33:12

would like this but I guess the

33:14

next thing I would would want to

33:16

see in in like Google Docs is

33:18

put it into suggest mode and just

33:20

do the thing where you like you

33:22

know you mark up all the changes

33:24

and then you can go through and

33:27

just like you're saying you can accept

33:29

or reject changes right. And then it

33:31

would also be interesting like when I

33:33

go and edit people's document right like.

33:35

you know I might make minor changes

33:37

but if I make a big change

33:40

to text I often add a comment

33:42

to that suggests change explaining what I'm

33:44

doing right and that would be interesting

33:46

if you had if you in Google

33:48

Docs it put it in suggest mode

33:50

and you know you've got the red

33:53

line and then the green part that's

33:55

like the new stuff yeah so maybe

33:57

someone the listeners can write in maybe

33:59

is there like a cursor version for

34:01

writing And what's the best version of

34:03

that? Like, maybe we just haven't found

34:06

what it is. There must be, given

34:08

so many AI companies out there, it

34:10

must exist somewhere. We just haven't seen

34:12

it. Yeah, yeah, but the Gemini stuff,

34:14

so it's so weird. I don't know.

34:16

I don't know what's going on with

34:18

it. It should be fantastic. But it

34:21

should be better. Definitely should be better.

34:23

Maybe we should. we should get a

34:25

notebook ML podcast where you get the

34:27

two hosts to talk about this phenomena

34:29

and you know we'll see see how

34:31

that comes out. We could play that

34:34

maybe as a maybe come December when

34:36

we need episodes we'll just slot that

34:38

in there. Well do you think that

34:40

they're like I mean I hate to

34:42

say this but there's probably a slot

34:44

podcast already out there based on that

34:47

format right? Oh, well, I haven't checked

34:49

in in a while, but back in

34:51

the fall, when it reached its zenith,

34:53

there was a D&D YouTube channel that

34:55

was just that. And they would do

34:57

AI-generated pictures, and then it would be

34:59

the two co-host talking with each other.

35:02

And they'd be like, this episode, strap

35:04

on your long-sword and gird your loins.

35:06

We're talking about goblins. That sounds so

35:08

horrible I want to get off the

35:10

internet. Like all things applying AI to

35:12

D&D, I think it was an early

35:15

way to kind of like do some

35:17

futurology on how you would apply this

35:19

stuff, right? You could kind of get

35:21

a sense of how that could apply

35:23

to kind of any domain that is

35:25

very well documented. Well, I think we

35:28

made some promise that we, I forget

35:30

if we weren't going to track it

35:32

or if we were going to see

35:34

major check-ins. But I think, I think

35:36

the, what do you call it? When

35:38

there is it a bellwether, an early

35:41

indicator, a touchstone, a leading indicator? Is

35:43

that what it is? I don't know.

35:45

A touchstone? You know, I think, I

35:47

think maybe, I think maybe, when we

35:49

look at the work from home, the

35:51

return to office thing, I think Dell

35:53

is the 2025 touchstone. It's like that

35:56

it's the indicator that we're going to

35:58

look at to see where the trends

36:00

are going. And I think kind of

36:02

like the chart that I put together

36:04

about which inning we're in. I think

36:06

Brandon has collected together what everyone's been

36:09

thinking this week, because apparently we're going

36:11

to work back. Well, let's work forward.

36:13

The year is, I guess it's a

36:15

month, but it's the date is sometime

36:17

in August 2020, and you basically I

36:19

think, what's his first name? Is it

36:22

Jim Clark? He basically, as I recall,

36:24

he used to work in QA in

36:26

the factory, eventually came up, he's more

36:28

or less number two at Dell, the

36:30

COO, and he said 60% of the

36:32

people who are working from home. They're

36:35

not going to be coming back to

36:37

the office regularly. Right, like, and I

36:39

was reminded of this back in the

36:41

COVID days, it was the sense of

36:43

like, huh, it works. I don't think

36:45

they're coming back to the office. Then

36:47

we fast forward March 2024, and this

36:50

was, I think we covered this at

36:52

the time, that Dell had this policy

36:54

of work in the office or you

36:56

don't get promoted, right? Like your career's

36:58

over, stagnates. And people were okay with

37:00

that. That was, that seemed to be

37:03

the feedback because everyone was like, like,

37:05

I was going to get a promotion

37:07

in the promotion in the first place

37:09

in the first place. Like it. I

37:11

got a paycheck. What more do I

37:13

need? And then this month, well, I

37:16

guess it's last month by five days.

37:18

It looks like they're totally entering the

37:20

hybrid thing and they're like, we had

37:22

a new deal for you, not only

37:24

you're not going to get promoted, fire

37:26

yourself if you don't want to come

37:28

into the office. I'm sure they got

37:31

a nice package and things like that.

37:33

But I think the other thing, I

37:35

think, I think. maybe this is not

37:37

in the storm of everything else not

37:39

so appreciated that this idea of return

37:41

to office in the tech world is

37:44

thoroughly infecting itself into the US federal

37:46

government. Lots of that stuff going on.

37:48

So I think we've got a we've

37:50

got a mainstreaming of this trend and

37:52

I and I got a feel that

37:54

like if we got Dell over here,

37:57

we got the federal government over here,

37:59

this is like this is some sort

38:01

of fashion transition. We're going from like

38:03

big pants to skinny pants, skinny pants

38:05

to big pants. Like I think we

38:07

might be at a tipping point. for

38:10

going either way. Now on the other

38:12

hand, if this whole new federal government...

38:14

situation doesn't work out well. It's I

38:16

think it's going to taint the return

38:18

to office are going to be like

38:20

listen. They made a return to offices

38:22

and the Hoover Dam burst. So obviously

38:25

yeah you got it you got to

38:27

work from home. I was about to

38:29

say like yeah RTO is one of

38:31

the smallest things that are going wrong

38:33

right now in the federal government so

38:35

like sure whatever you know you could

38:38

throw that in there but also like

38:40

the fact that maybe nobody will get

38:42

paid and you know Medicare is shut

38:44

down sure whatever you know RTO well

38:46

back on the Dell thing I did

38:48

think the one thing I always try

38:51

to think about like we're gonna talk

38:53

about like can we move the conversation

38:55

for it and I obviously know plenty

38:57

of friends that work in Dell just

38:59

being in Austin so obviously a lot

39:01

of people disappointed but I think you

39:04

know the way to kind of think

39:06

of it think of it is sort

39:08

of like and I'm stealing this phrase

39:10

from some other author it's not monetary

39:12

area not money in some ways, it's

39:14

just time wealth. And I think that's

39:16

really what comes down here. It's like

39:19

when you're forcing these people back into

39:21

the office, you basically are taking away

39:23

some time wealth from them. Basically, a

39:25

commute they didn't have to do a

39:27

little flexibility at their house, probably to

39:29

do some type of like taking care

39:32

of kids or picking up their kids

39:34

at school, and then you can even

39:36

go further. It's like, you know, what

39:38

is it worth to like? you know

39:40

as all parents kind of know to

39:42

often learn it's like as your kids

39:45

get older sometimes like that 15 minutes

39:47

in the cars like the best time

39:49

to be with your kids like you

39:51

know they'll kind of like tell you

39:53

what's going on and they're engaged and

39:55

they'll want to go on so it's

39:57

like there's kind of like a time

40:00

you know it's like that's kind of

40:02

what's like underlying all this it's sort

40:04

of like it's not just like oh

40:06

I have to go in the office

40:08

I mean conveniences you're actually 10% pay

40:10

cut or something matter, 20% pay cut.

40:13

Like, you'd really feel that, right? And

40:15

I think that's what a lot of

40:17

these people are feeling. It's like, oh,

40:19

I just taken this huge wealth cut.

40:21

And I think, I mean. It is

40:23

what it is. I mean, I don't

40:26

think these companies, I don't think companies

40:28

are gonna stop doing it, but it's,

40:30

it doesn't like really, it's hard, I

40:32

think, in companies, right, because you often

40:34

hear, like, you know, everyone, hey, come

40:36

to work, and, you know, we're here

40:39

for you, and, you know, we're here,

40:41

and, you know, we're, hey, come to

40:43

work, and, you know, we're here, we're,

40:45

we're, hey, come to work, and, and,

40:47

we're, we're, we're, we're, and, we're, we're,

40:49

we're, we're, we're, we're, we're, we're, we're,

40:51

and, we're, we're, we're, we're, we're, we're,

40:54

we're, and, we're, we're, we're, we're, we're,

40:56

we're, we're, we're, we're, we're, we're But

40:58

if you want, you can go to

41:00

the webinar, we're going to talk about

41:02

like how you can become and what

41:04

kind of exercises you can do to

41:07

like keep fit, right? And you're like,

41:09

oh, what I really like is just

41:11

be home taking a walk, you know?

41:13

So it is, it's a real, like,

41:15

it's a real, like, it's a real

41:17

struggle, I think, you know, between like,

41:20

what you're kind of exposed to at

41:22

work, right? Yeah, yeah, and you know

41:24

you you hit a we've said this

41:26

before but you the other thing that

41:28

you hit on is like it is

41:30

Something you were given and now it's

41:33

being taken away right with seemingly nothing

41:35

given in return Right, so and maybe

41:37

to be fair to the executive side

41:39

and think this is the other part

41:41

is like I think we just kind

41:43

of getting to it's like I don't

41:45

know, you know, you know, we'll call

41:48

executives just do not feel like they

41:50

can effectively manage without having more control

41:52

Like now we can debate whether that's

41:54

really true or not but like there's

41:56

clearly a feeling amongst executives that like

41:58

they just can't do it right you

42:01

know I mean I just cannot do

42:03

it so it's like so it's kind

42:05

of you know I get it and

42:07

I don't I don't even think it's

42:09

something it's like an opinion that can't

42:11

be it's like I like pizza like

42:14

you can't really debate it with someone

42:16

it's like once they've decided that it's

42:18

like well like what are you gonna

42:20

say like what are you gonna say?

42:22

The new thing also to add to

42:24

this is like there is some kind

42:26

of like there's something that drives these

42:29

executives crazy about people working from home.

42:31

It's just like, they must just wake

42:33

up with their jet lag in the

42:35

middle of the night and they're just

42:37

like, we got to do something about

42:39

this. Just somehow, you know, it kind

42:42

of amounts to like somehow they think

42:44

someone is taking their money, right? Like

42:46

there's something wrong with letting people work

42:48

from home. Like I guess they're not

42:50

productive enough. It's something must be driving

42:52

them batty. Because you know, we've seen

42:55

all of the papers that are just

42:57

like, doesn't really matter, right. And so.

42:59

Who knows? Just as they say on

43:01

the Oxide and Friends podcast, it's just

43:03

a dress code. Just, you know, choose

43:05

what you want to do. We'll see

43:08

how it pans out. Absolutely. Now you

43:10

mentioned, you mentioned a great article shared

43:12

by Richard Sarota, friend of the podcast,

43:14

from one of his favorites, Harvard Business

43:16

Review. HBR, as us insiders like to

43:18

call it, because, you know, we don't

43:20

like to spend a lot of time

43:23

saying the full name of things. We

43:25

got, you know. We got walks. We'd

43:27

rather spend minutes explaining the name. That's

43:29

right. And I think, you know, I

43:31

went back and reread this ahead of

43:33

the show and I missed the best

43:36

part. But let me give an overview.

43:38

It's kind of, it was saying like,

43:40

hey, you might be too positive at

43:42

work, especially if you're in management. You

43:44

might have toxic positivity where, you know,

43:46

you're always kind of, you're always trying

43:49

to point out the bright side of

43:51

things. you kind of move on to

43:53

like talking about how things are actually

43:55

good. And I think the first half

43:57

is like this sort of like multi-decade

43:59

realization that we're going through in personal

44:02

communications, especially on the, what would be

44:04

the proper phrasing, the identifying as male

44:06

side of the communication thing, which is

44:08

like sometimes, oftentimes, perhaps frequently times, when

44:10

people are expressing a negative emotion, they

44:12

just want you to be like... listen

44:14

to them and kind of help them

44:17

like be fine and you know I

44:19

think like you two I was raised

44:21

in the generation we're saying we probably

44:23

heard a lot said very sarcastically was

44:25

like what you need a hug right

44:27

and and I think the answer to

44:30

that is like yes everyone needs a

44:32

hug all the time it's great right

44:34

so like you know you can kind

44:36

of come from that perspective of like

44:38

it's good to not just be dismissive

44:40

of people like saying it's a bad

44:43

situation by saying it's actually great you

44:45

should not feel so poorly so you

44:47

don't want to have that you want

44:49

to deploy this this you know thing

44:51

of just listening to people and helping

44:53

them out but I think the best

44:55

part that I missed Very practical. I

44:58

like a practical tip is that if

45:00

you find yourself frequently posting positive things

45:02

on LinkedIn, you're probably doing it wrong.

45:04

And I wanted to run that by

45:06

you two in the form of a

45:08

question. How many times a day do

45:11

you think management can post something positive

45:13

on LinkedIn? That is, ooh, that is

45:15

a very difficult question. I don't know,

45:17

I think it's maybe less than a

45:19

day. Like, what's them once a day?

45:21

Oh, is the point three times? I'm

45:24

probably being to some kind of fraction.

45:26

I think it, well, I always think

45:28

like the LinkedIn stuff should be, if

45:30

you're gonna do it, it's something that

45:32

you think should be like very externalized,

45:34

right? So something that doesn't work that

45:37

your company will get a lot of

45:39

value of value of it, right. Right.

45:41

I hope that's number one. I hope

45:43

that's number one. I guess if you

45:45

have some incredible back-to-back moments that you

45:47

feel like you need to share, then

45:49

create. But it's, when you see people

45:52

doing it, you know, multiple times a

45:54

day, it does feel a little forced.

45:56

Which I think, which is this article,

45:58

that's like the one that really jumped

46:00

out at me is like, yeah, finally

46:02

someone, I just, it's like, it's like

46:05

everything that's good that's written, something, just

46:07

confirmation by us, like, oh, I've often

46:09

felt this way, like, like, like, like,

46:11

like, like, like, like, like, like, like,

46:13

like, like, like, like, like, like, I

46:15

mean back in the early days of

46:18

social media like Twitter 2010 or whatever

46:20

it would have been I felt like

46:22

I needed to read my entire feed

46:24

right you know oh what's happened the

46:26

last 24 hours I'll just go back

46:28

and I'll you know everything I missed

46:31

and you know eventually that slowly fell

46:33

off. LinkedIn has never felt like that.

46:35

LinkedIn has always just been like, what's

46:37

on? Oh God, yeah, there's just so

46:39

much noise. And I don't know why

46:41

so many people feel they need to

46:43

make more noise, but yeah, it's lost

46:46

a lot of signal. And I don't

46:48

know what the right answer is. I

46:50

don't know if LinkedIn is that. you

46:52

know platform or not but good lord

46:54

it's too noisy you know I don't

46:56

know if they have an API but

46:59

like that is like I think I

47:01

think I've I think I've I think

47:03

I've mentioned this over the years at

47:05

some point but there's there would be

47:07

a great client that I'm gonna tentatively

47:09

called the shit sprayer and and you

47:12

could put like a LinkedIn feed in

47:14

there maybe we're gonna use deep seek

47:16

so it's sufficient but you're basically just

47:18

gonna say hey A. because every now

47:20

and then there's something interesting in there

47:22

and just show me that like that

47:24

would be great if you could filter

47:27

feeds like that that you know there's

47:29

a new there's a new there's a

47:31

couple of new apps out I don't

47:33

even know what I'm looking for when

47:35

I log into late then I'm like

47:37

why I mean I want to see

47:40

like my friends are getting new jobs

47:42

as I say that that to me

47:44

is the number one thing the thing

47:46

that you want to know is like

47:48

when someone has moved and gone to

47:50

another company on to another company and

47:53

gotten to a new company and gotten

47:55

a new job for whatever company and

47:57

got or gotten a new job for

47:59

whatever job for whatever if you're linking

48:01

them you're probably just generally interested like

48:03

a social kind of thing like if

48:06

you see them next time you can

48:08

talk about two from a professional context

48:10

you're just like oh well they're at

48:12

this new place maybe that is something

48:14

maybe we can work together some way

48:16

because that's related to what I do

48:18

so like to me those like that's

48:21

the number one thing people are looking

48:23

for like everything else about like you

48:25

know the sales kickoff being great or

48:27

the marketing offsite was fantastic or you

48:29

know you know Thanks so and so

48:31

for being such a great. It's like,

48:34

okay, like that other stuff doesn't, I'm

48:36

just doesn't add. That feels like as

48:38

performative to use the word of the,

48:40

you know, I don't know, the. Here,

48:42

that's where it's like, yeah, I'd prefer

48:44

not to see it, because really nothing

48:47

for me to know about that. Yeah,

48:49

yeah, yeah, you know, maybe it is

48:51

just, the only thing I'll say is

48:53

like when I post content to LinkedIn,

48:55

it usually is the best performing content.

48:57

True, people are looking, that's a good

49:00

point. And that's why everyone should follow

49:02

the software defined talk, LinkedIn page, which

49:04

has nothing but good things that you

49:06

should read, like episodes that we're on

49:08

or podcast that we're on. Speaking of

49:10

nothing but good things, Brandon, do we

49:12

have any bureaucracy? Well, we did get

49:15

some listener feedback. I'm actually turn this

49:17

back over to, I'm going to start

49:19

with Matt. Matt, Warren Young wrote in,

49:21

and he had some comments about the

49:23

dot bash RC file and the dash

49:25

profile file. So I want to give

49:28

you an opportunity to clarify anything that

49:30

you may or may not have said

49:32

last week. I think I quickly over-generalized,

49:34

oh, one's interactive, one's not interactive. That's

49:36

not exactly right. We'll link to the

49:38

link and the show notes. There are

49:41

finer distinctions than that. We had our

49:43

friend chat GPT right up a little

49:45

differences, and we appreciate Warren for, you

49:47

know, kind of checking in with us

49:49

and explaining the difference. Yes. If you

49:51

need to know, you can go read

49:53

it. I looked at your chat GPT

49:56

thing and I thought I would dig

49:58

a little deeper, which I'm not going

50:00

to go over, because I wanted to

50:02

ask why. Why do you have this

50:04

stuff? And basically, one of them executes

50:06

when you kind of like log in

50:09

for the first time. I'm going to

50:11

mess it up. The other one also

50:13

executes whenever you have a brand new

50:15

shell that you do, or not. It's

50:17

very confusing. And it actually kind of

50:19

like made sense from an efficiency standpoint.

50:22

There's a certain set of things that

50:24

you only want to execute once for

50:26

all. shells or whatever you want to

50:28

call it that you have and so

50:30

I get it but the funny thing

50:32

about every time I asked chat GT

50:35

to do it more the last thing

50:37

you would say is commonly people source

50:39

the bash RC file and the bash

50:41

profile. So it was basically just like,

50:43

given all of that, people always just

50:45

execute both. Given all that, do both

50:47

is what the answer was. All right,

50:50

well, hopefully, it makes sense. I see

50:52

what you would have too. Hopefully that

50:54

clears up anything on the command line,

50:56

if you will. And then finally, on

50:58

a much lighter note, if you would

51:00

like a sticker, all you have to

51:03

do is send your postal address to

51:05

stickers at Software Fine Talk.com. I will

51:07

be happy to send you a sticker

51:09

anywhere in the world. Well, there's a

51:11

lot of conferences to go over. There's

51:13

CVO Navigate North America, February 10th and

51:16

11th in San Francisco. Devop stays LA

51:18

in scale 22X, which I'm not going

51:20

to be at, too bad. March 6th

51:22

to 9th in lovely, where is it,

51:24

Matt Ray? Sacramento, no, not Sacramento. It's

51:26

outside of El Pasadena. There you go.

51:29

I knew it was LA, not LA.

51:31

Very nice place there, Pasadena. You can

51:33

get a discount if you use the

51:35

code Devop. Singular. I'll be at the

51:37

V-Mug Netherlands March 12th. My my my

51:39

co-worker DeShan will be there too if

51:41

you enjoy DeShan. Devopsay Chicago is March

51:44

18th as well. I'll be at S-R-E-day

51:46

London March 27th and 28th and then

51:48

also the monkey grass the next day

51:50

March 27th and 28th in London. There's

51:52

Cloud Foundry Day US May 14th hosted

51:54

by us Broadcom via Muir Tanzoo. at

51:57

Apollo Alto offices May 14th in DC

51:59

Oslo where I'll be speaking May 21st

52:01

to 23rd and finally Kub Khan EU

52:03

in London April 1st to 4th. It

52:05

says good stuff. Speaking of good stuff,

52:07

Matt Ray what do you have to

52:10

recommend this episode? My good stuff is

52:12

You know I went and saw the

52:14

the flamey lips last weekend and my

52:16

recommendation is you know go out and

52:18

see more live music, but If you

52:20

know if they're coming to your town

52:22

go see them. It's you know It's

52:25

always one of the best shows I

52:27

see. So they go all out. There

52:29

was easily an inch deep of confetti

52:31

on the ground by the end of

52:33

it. Lights, confetti, fog machines. It was

52:35

insane. So highly recommended. Sounds like the

52:38

old days. It was like the old

52:40

days. Yeah, I was reminest over the

52:42

time I saw them and like. a

52:44

tiny tiny club and they turned on

52:46

their fog machines 10 months before they

52:48

started and you couldn't see anything. You

52:51

could barely breed there was so much

52:53

of that sugar fog in the air.

52:55

But it was an amazing show. They're

52:57

all amazing shows. Go see them. How

52:59

about yourself Brandon? When you're not bumping

53:01

into people in the sugar fog, or

53:04

you're not doing that, I've messed it

53:06

up. What do you recommend? Yeah, no

53:08

sugar fog here. What I'm recommending this

53:10

week is I got a blackstone griddle.

53:12

So I don't know, are you guys

53:14

griddle people or are you grill people?

53:16

Or do you not care? Not care.

53:19

Not care. Well, I have a griddle,

53:21

but it's for the stovetop. It's just

53:23

an old round cast iron thing that

53:25

I used to heat up tortillas and

53:27

things like that. Yeah, well I got

53:29

the, there's a link here, I got

53:32

the big griddle, and I replaced the

53:34

grill, so it's out on the deck,

53:36

and I actually quite like it. It

53:38

is great, it is surprisingly how convenient

53:40

it is to just have a giant

53:42

surface to just cook a bunch of

53:45

stuff. I've gone, I've taken and cooking,

53:47

instead of going to Japan to just

53:49

cook a bunch of stuff, I've gone,

53:51

I've taken to cooking, instead of going

53:53

to a griddle, we're giving away the

53:55

griddle. use the grill to grill. I

53:58

just kind of used it as like

54:00

a big cooking service. So I like

54:02

it. So if you're looking for another,

54:04

I don't know, outdoor grill, kitchen gadget,

54:06

I would recommend the Blackstone. So I

54:08

check that. And then of course you

54:10

can get into the watch all the

54:13

used two videos about how do you

54:15

season your griddle, which is like, you

54:17

know, that is a. If you will,

54:19

a rabbit hole, you can go down

54:21

that has no end. There's no end

54:23

to that. So if you like doing

54:26

that kind of stuff, get a griddle.

54:28

I like the cast iron and the

54:30

other carbon steel and I'm guessing griddle

54:32

people who are just like, there's two

54:34

types of people who will use as

54:36

people who wash it with soap and

54:39

people who don't. There's nothing better. I

54:41

was going to say it's about... Do

54:43

you use water and soap on cast

54:45

iron or do you not? You want

54:47

to see a motion? You want to

54:49

see something that brings the true person

54:51

out? Get yourself in a message board

54:54

about that. And then there's always the

54:56

one crazy person who was it with

54:58

salt and you're just like, I don't

55:00

know what's up with you. What are

55:02

you doing? Yeah, yeah, you know, I

55:04

haven't watched it in a while, but

55:07

you watch those those calming multi-hour videos

55:09

of short order cooks. using like an

55:11

industrial griddle, just looks like a lot

55:13

of fun, just frying up eggs like

55:15

there's no tomorrow. It's good stuff. Well,

55:17

I've already recommended, you know, I think

55:20

I think I understand the point of

55:22

canvas now. I've used it. I think

55:24

it's pretty good. But I've got two

55:26

other things, since we already talked about

55:28

that, like I said, I was in

55:30

Gint. And I always forget, even though

55:33

I eat them a lot, Belgian in

55:35

Belgium, are actually this whole other experience

55:37

that you don't cooked. cooked enough and

55:39

then somehow they put some sort of

55:41

sugary coating on it and then they

55:43

put it in the waffle iron again

55:45

making sure to line up the slots

55:48

and by doing that you get this

55:50

crispy sugary thing on the outside. Oh

55:52

so good. Like you're not going to

55:54

get that at Costco unfortunately so you

55:56

got to get it in Belgium. And

55:58

the other thing just a recommendation and

56:01

a thank you to Alex Williams from

56:03

Portland Oregon. Now he sent me this

56:05

zen. Remember zines called couches? And we'll

56:07

put a link to it in the

56:09

show notes, which is going to be

56:11

at software to find talk. I don't

56:14

think you can order it online. I

56:16

think you have to be in the

56:18

Portland area. But much like my kind

56:20

of winnowing obsession with garbage shares of

56:22

Amsterdam, this person went around Portland and

56:24

he found pictures of garbage couches and

56:27

he made a little zen of just

56:29

his pictures. of the garbage couches. So

56:31

you know as an officianado of this

56:33

of the the single seat version of

56:35

this I guess my version is very

56:37

European in that it's smaller the chair

56:39

and it suits just one person where

56:42

this is the American version it's a

56:44

big gigantic suburban equivalent of house furniture

56:46

but very good work from Christopher Michael

56:48

McMurray. Love that stuff. So check that

56:50

out if you're in the Portland organ

56:52

area. See if you can get your

56:55

hands on that scene. Now speaking of

56:57

things that you can get your hands

56:59

on, no matter where you are, whether

57:01

you're in Portland, Oregon, Portland, Maine, or

57:03

Portland everywhere on earth, you can always

57:05

listen to software to find talk. Now

57:08

you've already listened to it, or you've

57:10

already listened to it, or you've skipped

57:12

ahead for some reason, or maybe someone

57:14

in your family, a roommate is listening

57:16

to this out loud, and you're like,

57:18

thank God, this is almost better. You

57:20

know, nonetheless, you could always go to

57:23

software to find talk.com and figure out

57:25

how to subscribe. Or, you know, you

57:27

could say, why do you listen to

57:29

people joking about, you know, people murdering

57:31

people? That maybe doesn't seem like a

57:33

healthy way to live your life. No

57:36

more murders. Don't listen to a murder

57:38

podcast. Listen to software to find talk.com.

57:40

Now, there's a lot of things we

57:42

mentioned, including that couches thing, the griddle.

57:44

Other stuff if you want to get

57:46

all sorts of links you go to

57:49

software to find talk.com/five oh five You

57:51

can find those links and with that.

57:53

We'll see everyone next time. Bye. Bye

57:55

All right. Well, I just wanted I

57:57

throw it out only because I think

57:59

this is some of the funniest Dryas

58:02

and when I say that sort of

58:04

a very dry written analysis and that

58:06

was what he pasted it and out

58:08

so if you have not seen it

58:10

it's in the show notes you can

58:12

click on it and it's just a

58:14

funny little note about you know Tesla's

58:17

earnings in like and where the analyst

58:19

basically says there is no reason the

58:21

stock should have gone up and it

58:23

went up but he just says it

58:25

in a very funny way so it's

58:27

like a It's a frustrated liberal arts

58:30

major who's a great writer who's like,

58:32

you know what, I'm about to break

58:34

out some of my writing skills,

58:36

because why not? This is the time.

58:38

Yeah, I'm just going to read the

58:41

first line, because I think it is

58:43

maybe, I'm sure there, I don't read

58:45

a lot of financial on this stuff.

58:47

Maybe after all these years, there

58:49

is a better line, but this

58:51

is definitely, maybe the line of

58:54

the year. And it almost achieves,

58:56

It's not clear to us why

58:58

Tesla shares traded as much as

59:00

0.5% higher in the aftermarket Wednesday,

59:03

although we have some leading theories.

59:05

So I think when you have

59:08

the financial analyst at JPMC being

59:10

like, it's not clear to us.

59:12

Like they must have spent a lot

59:14

of time before they were like, we

59:16

don't know what the fuck's going on.

59:18

over here, you know, and, and, and,

59:20

or, or just as Benedict Evans pointed

59:23

out, he was like, I think that

59:25

was just a lot of words for

59:27

them to say, fuck this bullshit.